Bump up to 1.2.4 (#409)

Update README.rst

.travis.yml

@@ -5,7 +5,7 @@ python:
 - "3.4"  # 2.7 and 3.5 are preinstalled by default
 env:
   global:
-  - ETCDVERSION=3.0.15 ZKVERSION=3.4.9 CONSULVERSION=0.7.2
+  - ETCDVERSION=3.0.17 ZKVERSION=3.4.9 CONSULVERSION=0.7.4
   - PYVERSIONS="2.7 3.4 3.5"
   matrix:
   - TEST_SUITE="python setup.py"

README.rst

@@ -3,7 +3,7 @@
 Patroni: A Template for PostgreSQL HA with ZooKeeper, etcd or Consul
 ------------------------------------------------------------
 
-You can find a version of this documentation that is searchable and also easier to navigagate at `patroni.readthedocs.io <https://patroni.readthedocs.io>`__.
+You can find a version of this documentation that is searchable and also easier to navigate at `patroni.readthedocs.io <https://patroni.readthedocs.io>`__.
 
 
 There are many ways to run high availability with PostgreSQL; for a list, see the `PostgreSQL Documentation <https://wiki.postgresql.org/wiki/Replication,_Clustering,_and_Connection_Pooling>`__.
@@ -12,7 +12,7 @@ Patroni is a template for you to create your own customized, high-availability s
 
 We call Patroni a "template" because it is far from being a one-size-fits-all or plug-and-play replication system. It will have its own caveats. Use wisely.
 
-**Note to Kubernetes users**: We're currently developing Patroni to be as useful as possible for teams running Kubernetes on top of Google Compute Engine; Patroni can be the HA solution for Postgres in such an environment. To this end, we've created a `Helm Chart <https://github.com/kubernetes/charts/tree/master/incubator/patroni>`__ that enables you to deploy a five-node Patroni cluster using a Kubernetes PetSet.
+**Note to Kubernetes users**: We're currently developing Patroni to be as useful as possible for teams running Kubernetes on top of Google Compute Engine; Patroni can be the HA solution for Postgres in such an environment. To this end, there is a `Helm chart <https://github.com/kubernetes/charts/tree/master/incubator/patroni>`__ that uses Patroni and  `Spilo  <https://github.com/zalando/spilo/>`__ to provision a five-node PostgreSQL HA cluster in a Kubernetes+GCE environment. (The Helm chart deploys Spilo Docker images, not just "bare" Patroni.)
 
 .. contents::
     :local:
@@ -95,7 +95,7 @@ Go `here <https://github.com/zalando/patroni/blob/master/docs/ENVIRONMENT.rst>`_
 Replication Choices
 ===============
 
-Patroni uses Postgres' streaming replication, which is asynchronous by default. Patroni's asynchronous replication configuration allows for ``maximum_lag_on_failover`` settings. This setting ensures failover will not occur if a follower is more than a certain number of bytes behind the leader. This setting should be increased or decreased based on business requirements. It's also possible to use synchronous replication for better durability guarantees. See `replication modes documentation <https://github.com/zalando/patroni/blob/master/docs/replication_modes.rst>` for details.
+Patroni uses Postgres' streaming replication, which is asynchronous by default. Patroni's asynchronous replication configuration allows for ``maximum_lag_on_failover`` settings. This setting ensures failover will not occur if a follower is more than a certain number of bytes behind the leader. This setting should be increased or decreased based on business requirements. It's also possible to use synchronous replication for better durability guarantees. See `replication modes documentation <https://github.com/zalando/patroni/blob/master/docs/replication_modes.rst>`__ for details.
 
 ===============================
 Applications Should Not Use Superusers

patroni/dcs/etcd.py

@@ -530,6 +530,8 @@ class Etcd(AbstractDCS):
                 except etcd.EtcdWatchTimedOut:
                     self._client.http.clear()
                     return False
+                except (etcd.EtcdEventIndexCleared, etcd.EtcdWatcherCleared):  # Watch failed
+                    return True  # leave the loop, because watch with the same parameters will fail anyway
                 except etcd.EtcdException:
                     logger.exception('watch')
 

patroni/postgresql.py

@@ -177,6 +177,8 @@ class Postgresql(object):
                 parameters.pop('synchronous_standby_names', None)
             else:
                 parameters['synchronous_standby_names'] = self._synchronous_standby_names
+        if self._major_version >= 9.6 and parameters['wal_level'] == 'hot_standby':
+            parameters['wal_level'] = 'replica'
         return {k: v for k, v in parameters.items() if not self._major_version or
                 self._major_version >= self.CMDLINE_OPTIONS.get(k, (0, 1, 9.1))[2]}
 
@@ -249,8 +251,6 @@ class Postgresql(object):
                     elif r[0] in changes:
                         unit = changes['wal_segment_size'] if r[0] in ('min_wal_size', 'max_wal_size') else r[2]
                         new_value = changes.pop(r[0])
-                        if self._major_version >= 9.6 and r[0] == 'wal_level' and new_value == 'hot_standby':
-                            new_value = 'replica'
                         if new_value is None or not compare_values(r[3], unit, r[1], new_value):
                             if r[4] == 'postmaster':
                                 pending_restart = True
@@ -435,6 +435,7 @@ class Postgresql(object):
         if ret:
             self.write_pg_hba(config.get('pg_hba', []))
             self._major_version = self.get_major_version()
+            self._server_parameters = self.get_server_parameters(self.config)
         else:
             self.set_state('initdb failed')
         return ret
@@ -697,9 +698,7 @@ class Postgresql(object):
         self._write_postgresql_conf()
         self.resolve_connection_addresses()
 
-        opts = {p: self._server_parameters[p] for p, v in self.CMDLINE_OPTIONS.items() if self._major_version >= v[2]}
-        if self._major_version >= 9.6 and opts['wal_level'] == 'hot_standby':
-            opts['wal_level'] = 'replica'
+        opts = {p: self._server_parameters[p] for p in self.CMDLINE_OPTIONS if p in self._server_parameters}
         options = ['--{0}={1}'.format(p, v) for p, v in opts.items()]
 
         start_initiated = time.time()
@@ -1209,6 +1208,7 @@ $$""".format(name, ' '.join(options)), name, password, password)
         ret = self.create_replica(clone_member) == 0
         if ret:
             self._major_version = self.get_major_version()
+            self._server_parameters = self.get_server_parameters(self.config)
             self.delete_trigger_file()
             self.restore_configuration_files()
         return ret

patroni/scripts/wale_restore.py

@@ -133,8 +133,16 @@ class WALERestore(object):
                     with psycopg2.connect(self.master_connection) as con:
                         con.autocommit = True
                         with con.cursor() as cur:
-                            cur.execute("SELECT pg_xlog_location_diff(pg_current_xlog_location(), %s)",
-                                        (backup_start_lsn,))
+                            cur.execute("""SELECT CASE WHEN pg_is_in_recovery()
+                                                       THEN GREATEST(
+                                                                pg_xlog_location_diff(COALESCE(
+                                                                  pg_last_xlog_receive_location(), '0/0'), %s)::bigint,
+                                                                pg_xlog_location_diff(
+                                                                  pg_last_xlog_replay_location(), %s)::bigint)
+                                                       ELSE pg_xlog_location_diff(
+                                                                pg_current_xlog_location(), %s)::bigint
+                                                   END""", (backup_start_lsn, backup_start_lsn, backup_start_lsn))
+
                             diff_in_bytes = int(cur.fetchone()[0])
                 except psycopg2.Error:
                     logger.exception('could not determine difference with the master location')

patroni/version.py

@@ -1 +1 @@
-__version__ = '1.2.3'
+__version__ = '1.2.4'

tests/test_etcd.py

@@ -61,6 +61,8 @@ def etcd_watch(self, key, index=None, timeout=None, recursive=None):
         return etcd.EtcdResult('delete', {})
     elif timeout == 10.0:
         raise etcd.EtcdException
+    elif timeout == 20.0:
+        raise etcd.EtcdEventIndexCleared
 
 
 def etcd_write(self, key, value, **kwargs):
@@ -307,6 +309,7 @@ class TestEtcd(unittest.TestCase):
         self.etcd.watch(20729, 1.5)
         self.etcd.watch(20729, 4.5)
         with patch.object(AbstractDCS, 'watch', Mock()):
+            self.assertTrue(self.etcd.watch(20729, 19.5))
             self.etcd.watch(20729, 9.5)
 
     def test_other_exceptions(self):

tests/test_postgresql.py

@@ -41,8 +41,7 @@ class MockCursor(object):
                             ('search_path', 'public', None, 'string', 'user'),
                             ('port', '5433', None, 'integer', 'postmaster'),
                             ('listen_addresses', '*', None, 'string', 'postmaster'),
-                            ('autovacuum', 'on', None, 'bool', 'sighup'),
-                            ('wal_level', 'replica', None, 'enum', 'postmaster')]
+                            ('autovacuum', 'on', None, 'bool', 'sighup')]
         else:
             self.results = [(None, None, None, None, None, None, None, None, None, None)]
 
@@ -766,7 +765,7 @@ class TestPostgresql(unittest.TestCase):
         self.assertEquals(value_in_conf(), None)
 
     def test_get_server_parameters(self):
-        config = {'synchronous_mode': True, 'parameters': {}, 'listen': '0'}
+        config = {'synchronous_mode': True, 'parameters': {'wal_level': 'hot_standby'}, 'listen': '0'}
         self.p.get_server_parameters(config)
         self.p.set_synchronous_standby('foo')
         self.p.get_server_parameters(config)